The patterns of associations from analyses with a subsample of individuals aged 25 years or older (n = 2,738) were largely consistent with what we found with the whole sample (individuals aged 18 years or older) in terms of direction and magnitudes of associations (see Supplemental Table 1). However, several path coefficients (e.g., paternal bonding → offspring educational attainment; paternal education polygenic score → offspring educational attainment; parental education polygenic score → parental income) in the path models where parental education, income, and parenting were considered as mediators of the effects of parental education polygenic score on offspring educational attainment (Figure 3) became statistically non-significant, which was likely due to the reduction in sample size and statistical power. That the association between parental bonding and offspring educational attainment became non-significant in this subsample of offspring aged 25 years or older may also reflect developmental changes in parent-child relationships (Aquilino, 1997) and suggest a decline in the impact of parent-child relationship beyond emerging adulthood.
